I have this in my /etc/rc.conf: ifconfig_bridge0="inet 192.168.70.150/24 addm bge0 up" but when I do an ifconfig bridge0, I don't see an inet: entry for it

What am I missing or doing wrong?

mns: I'm guessing you also need something like cloned_interfaces="bridge0"

foxxx0: I have that too, sorry forgot to mentiont that.

mns: ifconfig_bridge0="inet 192.168.0.25 netmask 255.255.255.0 addm igb0 addm tap4 addm tap0 addm tap1 addm epair0a addm tap2 addm tap3"

maybe try adding the netmask and dropping the /24 ?

this is what i got for mine.. the "mess" are mainly for bhyve instances

i have ifconfig_bridge0="up" on line below

not sure why i did it that way but it works

meow there) I use DRBD quite a lot on gnu/linux for building convergent XEN clusters (storage+virtualization) and that's only possible thanks to DRBD being able to have dual-primaries during guest live-migration

as far as I've checked, HAST doesn't have dual-primaries. so I suppose XEN guest live-migration won't work there.

is it planned to have that feature for HAST?

(the moment where dual-primaries is used is very short, but seems necessary to me)

voy4g3r2: My problem is that although bridge0 is up, and the epair0a is up, and is a member, my vnet jail doesn't seem to be able to connect with the outside world. this used to work (famour last words, I know) .

rtprio: yup

polarian: so you either need to run nat on your wireguard host or run wireguard on your router instead

if i recall your problem

Hi, I'm struggling to configure /compat/ubuntu on FreeBSD 14.1

I installed is as described in docs.freebsd.org/en/books/handbook/linuxemu

Now, when I'm on chroot /compat/ubuntu /bin/bash

I cannot install basic apps, like curl for example.

is linux bash running now?

After "apt update" I run "apt install curl" and I get: pastebin.com/nns6Qa15

you already filled the sources.list ?

did you try running --fix-broken like it suggests?

rtprio: no, didn't filled sources.list, let me try

I anm trying to cheat a little to save a bit of time. trying to bring up a 4 drive raidz1 and build it in degraded mode while i wait for what will be the parity info drive to reformat from 520b sectors..

the plan was add the physical volumes da4-7 and then create a sparse zvol for a placeholder. create the raidz1 then offline the zvol.

the sparse zvol

id be replacing with the real deal in about 18 hours that its gonna take to let the drive reformat

"reformat from 512*b sectors" ?

from 520

they are sas drives from fibre channel arrays

uh. i don't think you can build a pool in degraded mode

why not just wait your 18 hours

youre right. you have to have it think all is there. its a bit of q sleight of hand trick.

Hi again, I was able to install some command line applications in /compat/ubuntu, now I installed an application that internally bundles a Java Runtime, but I'm getting: Desktop API is not suppoorted on the current platform.

I must add. This is running perfectly well on another FreeBSD 14.1 machine.

if you actually tried to fill a pool beyond the disk storage available for the sparse zvol it would crap itself

but once built you offline the drive that is the sparse zvol

the it runs in degraded

martinrame: i suppose if you had some extra space you could create the pool like this: truncate -s [exact size of drives] /usr/placeholder; zpool create blah raidz1 /usr/placeholder /dev/da4 /dev/da3 /dev/da2

one of those dragons be here things. :)

rtprio: this should be for mrelcee

I have the /compat/ubuntu issue, not the zfs pool one.

rrprio: exactly what i'm doing only with a sparse zvol. i unfortunately dont have 3.6tb free anywhere else atm.

to just use trunc

mrelcee: ok. worth a try i guess. I wouldn't expect it to work

martinrame: perhaps you missed a step on this system then the last one

specifying a zvol should be as simple as /dev/zvol/poolname/zvolname as the device. which i've done and I then used gpart to create a gpt label, and added a partition the exact size of the drives to that. so id be specifying it as /dev/zvol/poolname/zvolnamep1

To debug the issue I'm trying to run xcalc, but I get: /compat/ubuntu/bin/xcalc: /lib64/libm.so.6: version `GLIBC_2.29' not found (don't know how to install libm.so.6)

that should be part of glibc

rtprio: yes, I don't know why it doesn't exists here. I was able to run xclock, but not xcalc

probably because xclock doesn't link to libm

i mean you can simulate the same idea with disk files. add the three drives and then the disk file as a zpool. it'll limit the size to the size of the disk file that is added to it. you can offine that but it wont expand till theres a device listed in its place that is the proper size. hence i was going for a sparse zvol i just dont add data to. :)

mrelcee: ok man. but i would be prepared to rebuild the pool when you have all disks present

nah. replace the hinky device i the pool. let it resilver

ok, good luck

maybe i have a way. i do have 1T free ion my boot ssd.. and the data that needs on there is 1.8T. make a 950G file with truncate.. add that and the three drives.. move data over and when the last drive is done reformatting.. I can down the dummy file, replace with the real disk. let autoexpand do its job.. the sparse zvol not working is the problem here...

funny thing this worked in pre-openzfs days on freebsd. is advertised to work on linux machines now. doesnt now on freebsd. or I am missing something.

it seems like a lot of work to do it the wrong way first rather than just doing it the right way™️

just shifting time around. its called hacking. :) this isnt wrong like creating manbearpig. Just a little wrong.

yep creating 950G file and replacing the 10G placeholder it autoexpanded the volume.

that is offlined again - i was reformatting the drives yesterday and missed doing this oe. i really dont have another day to wait around is all. :)